[Search for users]
[Overall Top Noters]
[List of all Conferences]
[Download this site]
| Title: | DECmcc user notes file. Does not replace IPMT. |
| Notice: | Use IPMT for problems. Newsletter location in note 6187 |
| Moderator: | TAEC::BEROUD |
|
| Created: | Mon Aug 21 1989 |
| Last Modified: | Wed Jun 04 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 6497 |
| Total number of notes: | 27359 |
1889.0. "PM handling of %MCC-E-INSUF_BUF and memory leak" by SWORD1::ES (Eugene Shvartsman) Fri Dec 06 1991 13:05
When MM discovers that Out_P parameter is not big enough it sets the
mcc_w_curlen field of Out_P to required size, handle to state MORE and returns
status %MCC-E-INSUF_BUF. Seeing that PM calls MM again with increased Out_P.
If MM will use this Out_P everything would be fine.
So far so good, as should be.
But MM is no under any obligation to return the same information which it had
intended to return the first time. It may want more. So it does it again, i.e.
sets the mcc_w_curlen field of Out_P to required size, handle to state MORE
and returns status %MCC-E-INSUF_BUF. But this time we won't be so lucky.
PM will simply display %MCC-E-INSUF_BUF message and that will be it.
I may guess the intent of such behavior as to attempt to avoid infinite loop,
or whatever other reasons, but would you, please, issue a CANCEL request in this
case to allow the MM to clean memory.
Thank you,
Gene
| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 1889.1 | | VERNA::V_GILBERT | | Fri Dec 06 1991 14:20 | 3 |
| This has been entered as QAR 1791 for Iconic Map.
Verna
|
| 1889.2 | FCL to be fixed for v1.2 | TOOK::CALLANDER | MCC = My Constant Companion | Tue Jan 14 1992 10:47 | 7 |
| FCL has also been qared (thanks Keith) on this. Yes, the reasoning was
to avoid an infinite loop -- actually we hit the infinite loop and went
back and added the "do it only once". We are modifying the FCL to
attempt 3 times before giving up, and then passing a cancel down.
jill
|